ref: 9ae6dcb7e063d495267e7a53ddecb4d8356645ba
dir: /src/ext4_journal.c/
/*
* Copyright (c) 2015 Grzegorz Kostka (kostka.grzegorz@gmail.com)
* Copyright (c) 2015 Kaho Ng (ngkaho1234@gmail.com)
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
*
* - Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* - Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* - The name of the author may not be used to endorse or promote products
* derived from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
* IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
* OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
* IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
* THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
/** @addtogroup lwext4
* @{
*/
/**
* @file ext4_journal.c
* @brief Journal handle functions
*/
#include "ext4_config.h"
#include "ext4_types.h"
#include "ext4_misc.h"
#include "ext4_errno.h"
#include "ext4_debug.h"
#include "ext4_fs.h"
#include "ext4_super.h"
#include "ext4_journal.h"
#include "ext4_blockdev.h"
#include "ext4_crc32.h"
#include "ext4_journal.h"
#include <string.h>
#include <stdlib.h>
/**@brief Revoke entry during journal replay.*/
struct revoke_entry {
/**@brief Block number not to be replayed.*/
ext4_fsblk_t block;
/**@brief For any transaction id smaller
* than trans_id, records of @block
* in those transactions should not
* be replayed.*/
uint32_t trans_id;
/**@brief Revoke tree node.*/
RB_ENTRY(revoke_entry) revoke_node;
};
/**@brief Valid journal replay information.*/
struct recover_info {
/**@brief Starting transaction id.*/
uint32_t start_trans_id;
/**@brief Ending transaction id.*/
uint32_t last_trans_id;
/**@brief Used as internal argument.*/
uint32_t this_trans_id;
/**@brief No of transactions went through.*/
uint32_t trans_cnt;
/**@brief RB-Tree storing revoke entries.*/
RB_HEAD(jbd_revoke, revoke_entry) revoke_root;
};
/**@brief Journal replay internal arguments.*/
struct replay_arg {
/**@brief Journal replay information.*/
struct recover_info *info;
/**@brief Current block we are on.*/
uint32_t *this_block;
/**@brief Current trans_id we are on.*/
uint32_t this_trans_id;
};
/* Make sure we wrap around the log correctly! */
#define wrap(sb, var) \
do { \
if (var >= jbd_get32((sb), maxlen)) \
var -= (jbd_get32((sb), maxlen) - jbd_get32((sb), first)); \
} while (0)
static inline int32_t
trans_id_diff(uint32_t x, uint32_t y)
{
int32_t diff = x - y;
return diff;
}
static int
jbd_revoke_entry_cmp(struct revoke_entry *a, struct revoke_entry *b)
{
if (a->block > b->block)
return 1;
else if (a->block < b->block)
return -1;
return 0;
}
static int
jbd_block_rec_cmp(struct jbd_block_rec *a, struct jbd_block_rec *b)
{
if (a->lba > b->lba)
return 1;
else if (a->lba < b->lba)
return -1;
return 0;
}
static int
jbd_revoke_rec_cmp(struct jbd_revoke_rec *a, struct jbd_revoke_rec *b)
{
if (a->lba > b->lba)
return 1;
else if (a->lba < b->lba)
return -1;
return 0;
}
RB_GENERATE_INTERNAL(jbd_revoke, revoke_entry, revoke_node,
jbd_revoke_entry_cmp, static inline)
RB_GENERATE_INTERNAL(jbd_block, jbd_block_rec, block_rec_node,
jbd_block_rec_cmp, static inline)
RB_GENERATE_INTERNAL(jbd_revoke_tree, jbd_revoke_rec, revoke_node,
jbd_revoke_rec_cmp, static inline)
#define jbd_alloc_revoke_entry() ext4_calloc(1, sizeof(struct revoke_entry))
#define jbd_free_revoke_entry(addr) ext4_free(addr)
static int jbd_has_csum(struct jbd_sb *jbd_sb)
{
if (JBD_HAS_INCOMPAT_FEATURE(jbd_sb, JBD_FEATURE_INCOMPAT_CSUM_V2))
return 2;
if (JBD_HAS_INCOMPAT_FEATURE(jbd_sb, JBD_FEATURE_INCOMPAT_CSUM_V3))
return 3;
return 0;
}
#if CONFIG_META_CSUM_ENABLE
static uint32_t jbd_sb_csum(struct jbd_sb *jbd_sb)
{
uint32_t checksum = 0;
if (jbd_has_csum(jbd_sb)) {
uint32_t orig_checksum = jbd_sb->checksum;
jbd_set32(jbd_sb, checksum, 0);
/* Calculate crc32c checksum against tho whole superblock */
checksum = ext4_crc32c(EXT4_CRC32_INIT, jbd_sb,
JBD_SUPERBLOCK_SIZE);
jbd_sb->checksum = orig_checksum;
}
return checksum;
}
#else
#define jbd_sb_csum(...) 0
#endif
static void jbd_sb_csum_set(struct jbd_sb *jbd_sb)
{
if (!jbd_has_csum(jbd_sb))
return;
jbd_set32(jbd_sb, checksum, jbd_sb_csum(jbd_sb));
}
#if CONFIG_META_CSUM_ENABLE
static bool
jbd_verify_sb_csum(struct jbd_sb *jbd_sb)
{
if (!jbd_has_csum(jbd_sb))
return true;
return jbd_sb_csum(jbd_sb) == jbd_get32(jbd_sb, checksum);
}
#else
#define jbd_verify_sb_csum(...) true
#endif
#if CONFIG_META_CSUM_ENABLE
static uint32_t jbd_meta_csum(struct jbd_fs *jbd_fs,
struct jbd_bhdr *bhdr)
{
uint32_t checksum = 0;
if (jbd_has_csum(&jbd_fs->sb)) {
uint32_t block_size = jbd_get32(&jbd_fs->sb, blocksize);
struct jbd_block_tail *tail =
(struct jbd_block_tail *)((char *)bhdr + block_size -
sizeof(struct jbd_block_tail));
uint32_t orig_checksum = tail->checksum;
tail->checksum = 0;
/* First calculate crc32c checksum against fs uuid */
checksum = ext4_crc32c(EXT4_CRC32_INIT, jbd_fs->sb.uuid,
sizeof(jbd_fs->sb.uuid));
/* Calculate crc32c checksum against tho whole block */
checksum = ext4_crc32c(checksum, bhdr,
block_size);
tail->checksum = orig_checksum;
}
return checksum;
}
#else
#define jbd_meta_csum(...) 0
#endif
static void jbd_meta_csum_set(struct jbd_fs *jbd_fs,
struct jbd_bhdr *bhdr)
{
uint32_t block_size = jbd_get32(&jbd_fs->sb, blocksize);
struct jbd_block_tail *tail = (struct jbd_block_tail *)
((char *)bhdr + block_size -
sizeof(struct jbd_block_tail));
if (!jbd_has_csum(&jbd_fs->sb))
return;
tail->checksum = to_be32(jbd_meta_csum(jbd_fs, bhdr));
}
#if CONFIG_META_CSUM_ENABLE
static bool
jbd_verify_meta_csum(struct jbd_fs *jbd_fs,
struct jbd_bhdr *bhdr)
{
uint32_t block_size = jbd_get32(&jbd_fs->sb, blocksize);
struct jbd_block_tail *tail = (struct jbd_block_tail *)
((char *)bhdr + block_size -
sizeof(struct jbd_block_tail));
if (!jbd_has_csum(&jbd_fs->sb))
return true;
return jbd_meta_csum(jbd_fs, bhdr) == to_be32(tail->checksum);
}
#else
#define jbd_verify_meta_csum(...) true
#endif
#if CONFIG_META_CSUM_ENABLE
static uint32_t jbd_commit_csum(struct jbd_fs *jbd_fs,
struct jbd_commit_header *header)
{
uint32_t checksum = 0;
if (jbd_has_csum(&jbd_fs->sb)) {
uint32_t orig_checksum_type = header->chksum_type,
orig_checksum_size = header->chksum_size,
orig_checksum = header->chksum[0];
uint32_t block_size = jbd_get32(&jbd_fs->sb, blocksize);
header->chksum_type = 0;
header->chksum_size = 0;
header->chksum[0] = 0;
/* First calculate crc32c checksum against fs uuid */
checksum = ext4_crc32c(EXT4_CRC32_INIT, jbd_fs->sb.uuid,
sizeof(jbd_fs->sb.uuid));
/* Calculate crc32c checksum against tho whole block */
checksum = ext4_crc32c(checksum, header,
block_size);
header->chksum_type = orig_checksum_type;
header->chksum_size = orig_checksum_size;
header->chksum[0] = orig_checksum;
}
return checksum;
}
#else
#define jbd_commit_csum(...) 0
#endif
static void jbd_commit_csum_set(struct jbd_fs *jbd_fs,
struct jbd_commit_header *header)
{
if (!jbd_has_csum(&jbd_fs->sb))
return;
header->chksum_type = 0;
header->chksum_size = 0;
header->chksum[0] = jbd_commit_csum(jbd_fs, header);
}
#if CONFIG_META_CSUM_ENABLE
static bool jbd_verify_commit_csum(struct jbd_fs *jbd_fs,
struct jbd_commit_header *header)
{
if (!jbd_has_csum(&jbd_fs->sb))
return true;
return header->chksum[0] == to_be32(jbd_commit_csum(jbd_fs,
header));
}
#else
#define jbd_verify_commit_csum(...) true
#endif
#if CONFIG_META_CSUM_ENABLE
/*
* NOTE: We only make use of @csum parameter when
* JBD_FEATURE_COMPAT_CHECKSUM is enabled.
*/
static uint32_t jbd_block_csum(struct jbd_fs *jbd_fs, const void *buf,
uint32_t csum,
uint32_t sequence)
{
uint32_t checksum = 0;
if (jbd_has_csum(&jbd_fs->sb)) {
uint32_t block_size = jbd_get32(&jbd_fs->sb, blocksize);
/* First calculate crc32c checksum against fs uuid */
checksum = ext4_crc32c(EXT4_CRC32_INIT, jbd_fs->sb.uuid,
sizeof(jbd_fs->sb.uuid));
/* Then calculate crc32c checksum against sequence no. */
checksum = ext4_crc32c(checksum, &sequence,
sizeof(uint32_t));
/* Calculate crc32c checksum against tho whole block */
checksum = ext4_crc32c(checksum, buf,
block_size);
} else if (JBD_HAS_INCOMPAT_FEATURE(&jbd_fs->sb,
JBD_FEATURE_COMPAT_CHECKSUM)) {
uint32_t block_size = jbd_get32(&jbd_fs->sb, blocksize);
/* Calculate crc32c checksum against tho whole block */
checksum = ext4_crc32(csum, buf,
block_size);
}
return checksum;
}
#else
#define jbd_block_csum(...) 0
#endif
static void jbd_block_tag_csum_set(struct jbd_fs *jbd_fs, void *__tag,
uint32_t checksum)
{
int ver = jbd_has_csum(&jbd_fs->sb);
if (!ver)
return;
if (ver == 2) {
struct jbd_block_tag *tag = __tag;
tag->checksum = (uint16_t)to_be32(checksum);
} else {
struct jbd_block_tag3 *tag = __tag;
tag->checksum = to_be32(checksum);
}
}
/**@brief Write jbd superblock to disk.
* @param jbd_fs jbd filesystem
* @param s jbd superblock
* @return standard error code*/
static int jbd_sb_write(struct jbd_fs *jbd_fs, struct jbd_sb *s)
{
int rc;
struct ext4_fs *fs = jbd_fs->inode_ref.fs;
uint64_t offset;
ext4_fsblk_t fblock;
rc = jbd_inode_bmap(jbd_fs, 0, &fblock);
if (rc != EOK)
return rc;
jbd_sb_csum_set(s);
offset = fblock * ext4_sb_get_block_size(&fs->sb);
return ext4_block_writebytes(fs->bdev, offset, s,
EXT4_SUPERBLOCK_SIZE);
}
/**@brief Read jbd superblock from disk.
* @param jbd_fs jbd filesystem
* @param s jbd superblock
* @return standard error code*/
static int jbd_sb_read(struct jbd_fs *jbd_fs, struct jbd_sb *s)
{
int rc;
struct ext4_fs *fs = jbd_fs->inode_ref.fs;
uint64_t offset;
ext4_fsblk_t fblock;
rc = jbd_inode_bmap(jbd_fs, 0, &fblock);
if (rc != EOK)
return rc;
offset = fblock * ext4_sb_get_block_size(&fs->sb);
return ext4_block_readbytes(fs->bdev, offset, s,
EXT4_SUPERBLOCK_SIZE);
}
/**@brief Verify jbd superblock.
* @param sb jbd superblock
* @return true if jbd superblock is valid */
static bool jbd_verify_sb(struct jbd_sb *sb)
{
struct jbd_bhdr *header = &sb->header;
if (jbd_get32(header, magic) != JBD_MAGIC_NUMBER)
return false;
if (jbd_get32(header, blocktype) != JBD_SUPERBLOCK &&
jbd_get32(header, blocktype) != JBD_SUPERBLOCK_V2)
return false;
return jbd_verify_sb_csum(sb);
}
/**@brief Write back dirty jbd superblock to disk.
* @param jbd_fs jbd filesystem
* @return standard error code*/
static int jbd_write_sb(struct jbd_fs *jbd_fs)
{
int rc = EOK;
if (jbd_fs->dirty) {
rc = jbd_sb_write(jbd_fs, &jbd_fs->sb);
if (rc != EOK)
return rc;
jbd_fs->dirty = false;
}
return rc;
}
/**@brief Get reference to jbd filesystem.
* @param fs Filesystem to load journal of
* @param jbd_fs jbd filesystem
* @return standard error code*/
int jbd_get_fs(struct ext4_fs *fs,
struct jbd_fs *jbd_fs)
{
int rc;
uint32_t journal_ino;
memset(jbd_fs, 0, sizeof(struct jbd_fs));
/* See if there is journal inode on this filesystem.*/
/* FIXME: detection on existance ofbkejournal bdev is
* missing.*/
journal_ino = ext4_get32(&fs->sb, journal_inode_number);
rc = ext4_fs_get_inode_ref(fs,
journal_ino,
&jbd_fs->inode_ref);
if (rc != EOK) {
memset(jbd_fs, 0, sizeof(struct jbd_fs));
return rc;
}
rc = jbd_sb_read(jbd_fs, &jbd_fs->sb);
if (rc != EOK) {
memset(jbd_fs, 0, sizeof(struct jbd_fs));
ext4_fs_put_inode_ref(&jbd_fs->inode_ref);
return rc;
}
if (!jbd_verify_sb(&jbd_fs->sb)) {
memset(jbd_fs, 0, sizeof(struct jbd_fs));
ext4_fs_put_inode_ref(&jbd_fs->inode_ref);
rc = EIO;
}
if (rc == EOK)
jbd_fs->bdev = fs->bdev;
return rc;
}
/**@brief Put reference of jbd filesystem.
* @param jbd_fs jbd filesystem
* @return standard error code*/
int jbd_put_fs(struct jbd_fs *jbd_fs)
{
int rc = EOK;
rc = jbd_write_sb(jbd_fs);
ext4_fs_put_inode_ref(&jbd_fs->inode_ref);
return rc;
}
/**@brief Data block lookup helper.
* @param jbd_fs jbd filesystem
* @param iblock block index
* @param fblock logical block address
* @return standard error code*/
int jbd_inode_bmap(struct jbd_fs *jbd_fs,
ext4_lblk_t iblock,
ext4_fsblk_t *fblock)
{
int rc = ext4_fs_get_inode_dblk_idx(
&jbd_fs->inode_ref,
iblock,
fblock,
false);
return rc;
}
/**@brief jbd block get function (through cache).
* @param jbd_fs jbd filesystem
* @param block block descriptor
* @param fblock jbd logical block address
* @return standard error code*/
static int jbd_block_get(struct jbd_fs *jbd_fs,
struct ext4_block *block,
ext4_fsblk_t fblock)
{
/* TODO: journal device. */
int rc;
struct ext4_blockdev *bdev = jbd_fs->bdev;
ext4_lblk_t iblock = (ext4_lblk_t)fblock;
/* Lookup the logical block address of
* fblock.*/
rc = jbd_inode_bmap(jbd_fs, iblock,
&fblock);
if (rc != EOK)
return rc;
rc = ext4_block_get(bdev, block, fblock);
/* If succeeded, mark buffer as BC_FLUSH to indicate
* that data should be written to disk immediately.*/
if (rc == EOK) {
ext4_bcache_set_flag(block->buf, BC_FLUSH);
/* As we don't want to occupy too much space
* in block cache, we set this buffer BC_TMP.*/
ext4_bcache_set_flag(block->buf, BC_TMP);
}
return rc;
}
/**@brief jbd block get function (through cache, don't read).
* @param jbd_fs jbd filesystem
* @param block block descriptor
* @param fblock jbd logical block address
* @return standard error code*/
static int jbd_block_get_noread(struct jbd_fs *jbd_fs,
struct ext4_block *block,
ext4_fsblk_t fblock)
{
/* TODO: journal device. */
int rc;
struct ext4_blockdev *bdev = jbd_fs->bdev;
ext4_lblk_t iblock = (ext4_lblk_t)fblock;
rc = jbd_inode_bmap(jbd_fs, iblock,
&fblock);
if (rc != EOK)
return rc;
rc = ext4_block_get_noread(bdev, block, fblock);
if (rc == EOK)
ext4_bcache_set_flag(block->buf, BC_FLUSH);
return rc;
}
/**@brief jbd block set procedure (through cache).
* @param jbd_fs jbd filesystem
* @param block block descriptor
* @return standard error code*/
static int jbd_block_set(struct jbd_fs *jbd_fs,
struct ext4_block *block)
{
struct ext4_blockdev *bdev = jbd_fs->bdev;
return ext4_block_set(bdev, block);
}
/**@brief helper functions to calculate
* block tag size, not including UUID part.
* @param jbd_fs jbd filesystem
* @return tag size in bytes*/
static int jbd_tag_bytes(struct jbd_fs *jbd_fs)
{
int size;
/* It is very easy to deal with the case which
* JBD_FEATURE_INCOMPAT_CSUM_V3 is enabled.*/
if (JBD_HAS_INCOMPAT_FEATURE(&jbd_fs->sb,
JBD_FEATURE_INCOMPAT_CSUM_V3))
return sizeof(struct jbd_block_tag3);
size = sizeof(struct jbd_block_tag);
/* If JBD_FEATURE_INCOMPAT_CSUM_V2 is enabled,
* add 2 bytes to size.*/
if (JBD_HAS_INCOMPAT_FEATURE(&jbd_fs->sb,
JBD_FEATURE_INCOMPAT_CSUM_V2))
size += sizeof(uint16_t);
if (JBD_HAS_INCOMPAT_FEATURE(&jbd_fs->sb,
JBD_FEATURE_INCOMPAT_64BIT))
return size;
/* If block number is 4 bytes in size,
* minus 4 bytes from size */
return size - sizeof(uint32_t);
}
/**@brief Tag information. */
struct tag_info {
/**@brief Tag size in bytes, including UUID part.*/
int tag_bytes;
/**@brief block number stored in this tag.*/
ext4_fsblk_t block;
/**@brief Is the first 4 bytes of block equals to
* JBD_MAGIC_NUMBER? */
bool is_escape;
/**@brief whether UUID part exists or not.*/
bool uuid_exist;
/**@brief UUID content if UUID part exists.*/
uint8_t uuid[UUID_SIZE];
/**@brief Is this the last tag? */
bool last_tag;
/**@brief crc32c checksum. */
uint32_t checksum;
};
/**@brief Extract information from a block tag.
* @param __tag pointer to the block tag
* @param tag_bytes block tag size of this jbd filesystem
* @param remaining size in buffer containing the block tag
* @param tag_info information of this tag.
* @return EOK when succeed, otherwise return EINVAL.*/
static int
jbd_extract_block_tag(struct jbd_fs *jbd_fs,
void *__tag,
int tag_bytes,
int32_t remain_buf_size,
struct tag_info *tag_info)
{
char *uuid_start;
tag_info->tag_bytes = tag_bytes;
tag_info->uuid_exist = false;
tag_info->last_tag = false;
tag_info->is_escape = false;
/* See whether it is possible to hold a valid block tag.*/
if (remain_buf_size - tag_bytes < 0)
return EINVAL;
if (JBD_HAS_INCOMPAT_FEATURE(&jbd_fs->sb,
JBD_FEATURE_INCOMPAT_CSUM_V3)) {
struct jbd_block_tag3 *tag = __tag;
tag_info->block = jbd_get32(tag, blocknr);
if (JBD_HAS_INCOMPAT_FEATURE(&jbd_fs->sb,
JBD_FEATURE_INCOMPAT_64BIT))
tag_info->block |=
(uint64_t)jbd_get32(tag, blocknr_high) << 32;
if (jbd_get32(tag, flags) & JBD_FLAG_ESCAPE)
tag_info->is_escape = true;
if (!(jbd_get32(tag, flags) & JBD_FLAG_SAME_UUID)) {
/* See whether it is possible to hold UUID part.*/
if (remain_buf_size - tag_bytes < UUID_SIZE)
return EINVAL;
uuid_start = (char *)tag + tag_bytes;
tag_info->uuid_exist = true;
tag_info->tag_bytes += UUID_SIZE;
memcpy(tag_info->uuid, uuid_start, UUID_SIZE);
}
if (jbd_get32(tag, flags) & JBD_FLAG_LAST_TAG)
tag_info->last_tag = true;
} else {
struct jbd_block_tag *tag = __tag;
tag_info->block = jbd_get32(tag, blocknr);
if (JBD_HAS_INCOMPAT_FEATURE(&jbd_fs->sb,
JBD_FEATURE_INCOMPAT_64BIT))
tag_info->block |=
(uint64_t)jbd_get32(tag, blocknr_high) << 32;
if (jbd_get16(tag, flags) & JBD_FLAG_ESCAPE)
tag_info->is_escape = true;
if (!(jbd_get16(tag, flags) & JBD_FLAG_SAME_UUID)) {
/* See whether it is possible to hold UUID part.*/
if (remain_buf_size - tag_bytes < UUID_SIZE)
return EINVAL;
uuid_start = (char *)tag + tag_bytes;
tag_info->uuid_exist = true;
tag_info->tag_bytes += UUID_SIZE;
memcpy(tag_info->uuid, uuid_start, UUID_SIZE);
}
if (jbd_get16(tag, flags) & JBD_FLAG_LAST_TAG)
tag_info->last_tag = true;
}
return EOK;
}
/**@brief Write information to a block tag.
* @param __tag pointer to the block tag
* @param remaining size in buffer containing the block tag
* @param tag_info information of this tag.
* @return EOK when succeed, otherwise return EINVAL.*/
static int
jbd_write_block_tag(struct jbd_fs *jbd_fs,
void *__tag,
int32_t remain_buf_size,
struct tag_info *tag_info)
{
char *uuid_start;
int tag_bytes = jbd_tag_bytes(jbd_fs);
tag_info->tag_bytes = tag_bytes;
/* See whether it is possible to hold a valid block tag.*/
if (remain_buf_size - tag_bytes < 0)
return EINVAL;
if (JBD_HAS_INCOMPAT_FEATURE(&jbd_fs->sb,
JBD_FEATURE_INCOMPAT_CSUM_V3)) {
struct jbd_block_tag3 *tag = __tag;
memset(tag, 0, sizeof(struct jbd_block_tag3));
jbd_set32(tag, blocknr, (uint32_t)tag_info->block);
if (JBD_HAS_INCOMPAT_FEATURE(&jbd_fs->sb,
JBD_FEATURE_INCOMPAT_64BIT))
jbd_set32(tag, blocknr_high, tag_info->block >> 32);
if (tag_info->uuid_exist) {
/* See whether it is possible to hold UUID part.*/
if (remain_buf_size - tag_bytes < UUID_SIZE)
return EINVAL;
uuid_start = (char *)tag + tag_bytes;
tag_info->tag_bytes += UUID_SIZE;
memcpy(uuid_start, tag_info->uuid, UUID_SIZE);
} else
jbd_set32(tag, flags,
jbd_get32(tag, flags) | JBD_FLAG_SAME_UUID);
jbd_block_tag_csum_set(jbd_fs, __tag, tag_info->checksum);
if (tag_info->last_tag)
jbd_set32(tag, flags,
jbd_get32(tag, flags) | JBD_FLAG_LAST_TAG);
if (tag_info->is_escape)
jbd_set32(tag, flags,
jbd_get32(tag, flags) | JBD_FLAG_ESCAPE);
} else {
struct jbd_block_tag *tag = __tag;
memset(tag, 0, sizeof(struct jbd_block_tag));
jbd_set32(tag, blocknr, (uint32_t)tag_info->block);
if (JBD_HAS_INCOMPAT_FEATURE(&jbd_fs->sb,
JBD_FEATURE_INCOMPAT_64BIT))
jbd_set32(tag, blocknr_high, tag_info->block >> 32);
if (tag_info->uuid_exist) {
/* See whether it is possible to hold UUID part.*/
if (remain_buf_size - tag_bytes < UUID_SIZE)
return EINVAL;
uuid_start = (char *)tag + tag_bytes;
tag_info->tag_bytes += UUID_SIZE;
memcpy(uuid_start, tag_info->uuid, UUID_SIZE);
} else
jbd_set16(tag, flags,
jbd_get16(tag, flags) | JBD_FLAG_SAME_UUID);
jbd_block_tag_csum_set(jbd_fs, __tag, tag_info->checksum);
if (tag_info->last_tag)
jbd_set16(tag, flags,
jbd_get16(tag, flags) | JBD_FLAG_LAST_TAG);
if (tag_info->is_escape)
jbd_set16(tag, flags,
jbd_get16(tag, flags) | JBD_FLAG_ESCAPE);
}
return EOK;
}
/**@brief Iterate all block tags in a block.
* @param jbd_fs jbd filesystem
* @param __tag_start pointer to the block
* @param tag_tbl_size size of the block
* @param func callback routine to indicate that
* a block tag is found
* @param arg additional argument to be passed to func */
static void
jbd_iterate_block_table(struct jbd_fs *jbd_fs,
void *__tag_start,
int32_t tag_tbl_size,
void (*func)(struct jbd_fs * jbd_fs,
struct tag_info *tag_info,
void *arg),
void *arg)
{
char *tag_start, *tag_ptr;
int tag_bytes = jbd_tag_bytes(jbd_fs);
tag_start = __tag_start;
tag_ptr = tag_start;
/* Cut off the size of block tail storing checksum. */
if (JBD_HAS_INCOMPAT_FEATURE(&jbd_fs->sb,
JBD_FEATURE_INCOMPAT_CSUM_V2) ||
JBD_HAS_INCOMPAT_FEATURE(&jbd_fs->sb,
JBD_FEATURE_INCOMPAT_CSUM_V3))
tag_tbl_size -= sizeof(struct jbd_block_tail);
while (tag_tbl_size) {
struct tag_info tag_info;
int rc = jbd_extract_block_tag(jbd_fs,
tag_ptr,
tag_bytes,
tag_tbl_size,
&tag_info);
if (rc != EOK)
break;
if (func)
func(jbd_fs, &tag_info, arg);
/* Stop the iteration when we reach the last tag. */
if (tag_info.last_tag)
break;
tag_ptr += tag_info.tag_bytes;
tag_tbl_size -= tag_info.tag_bytes;
}
}
static void jbd_display_block_tags(struct jbd_fs *jbd_fs,
struct tag_info *tag_info,
void *arg)
{
uint32_t *iblock = arg;
ext4_dbg(DEBUG_JBD, "Block in block_tag: %" PRIu64 "\n", tag_info->block);
(*iblock)++;
wrap(&jbd_fs->sb, *iblock);
(void)jbd_fs;
return;
}
static struct revoke_entry *
jbd_revoke_entry_lookup(struct recover_info *info, ext4_fsblk_t block)
{
struct revoke_entry tmp = {
.block = block
};
return RB_FIND(jbd_revoke, &info->revoke_root, &tmp);
}
/**@brief Replay a block in a transaction.
* @param jbd_fs jbd filesystem
* @param tag_info tag_info of the logged block.*/
static void jbd_replay_block_tags(struct jbd_fs *jbd_fs,
struct tag_info *tag_info,
void *__arg)
{
int r;
struct replay_arg *arg = __arg;
struct recover_info *info = arg->info;
uint32_t *this_block = arg->this_block;
struct revoke_entry *revoke_entry;
struct ext4_block journal_block, ext4_block;
struct ext4_fs *fs = jbd_fs->inode_ref.fs;
(*this_block)++;
wrap(&jbd_fs->sb, *this_block);
/* We replay this block only if the current transaction id
* is equal or greater than that in revoke entry.*/
revoke_entry = jbd_revoke_entry_lookup(info, tag_info->block);
if (revoke_entry &&
trans_id_diff(arg->this_trans_id, revoke_entry->trans_id) <= 0)
return;
ext4_dbg(DEBUG_JBD,
"Replaying block in block_tag: %" PRIu64 "\n",
tag_info->block);
r = jbd_block_get(jbd_fs, &journal_block, *this_block);
if (r != EOK)
return;
/* We need special treatment for ext4 superblock. */
if (tag_info->block) {
r = ext4_block_get_noread(fs->bdev, &ext4_block, tag_info->block);
if (r != EOK) {
jbd_block_set(jbd_fs, &journal_block);
return;
}
memcpy(ext4_block.data,
journal_block.data,
jbd_get32(&jbd_fs->sb, blocksize));
if (tag_info->is_escape)
((struct jbd_bhdr *)ext4_block.data)->magic =
to_be32(JBD_MAGIC_NUMBER);
ext4_bcache_set_dirty(ext4_block.buf);
ext4_block_set(fs->bdev, &ext4_block);
} else {
uint16_t mount_count, state;
mount_count = ext4_get16(&fs->sb, mount_count);
state = ext4_get16(&fs->sb, state);
memcpy(&fs->sb,
journal_block.data + EXT4_SUPERBLOCK_OFFSET,
EXT4_SUPERBLOCK_SIZE);
/* Mark system as mounted */
ext4_set16(&fs->sb, state, state);
r = ext4_sb_write(fs->bdev, &fs->sb);
if (r != EOK)
return;
/*Update mount count*/
ext4_set16(&fs->sb, mount_count, mount_count);
}
jbd_block_set(jbd_fs, &journal_block);
return;
}
/**@brief Add block address to revoke tree, along with
* its transaction id.
* @param info journal replay info
* @param block block address to be replayed.*/
static void jbd_add_revoke_block_tags(struct recover_info *info,
ext4_fsblk_t block)
{
struct revoke_entry *revoke_entry;
ext4_dbg(DEBUG_JBD, "Add block %" PRIu64 " to revoke tree\n", block);
/* If the revoke entry with respect to the block address
* exists already, update its transaction id.*/
revoke_entry = jbd_revoke_entry_lookup(info, block);
if (revoke_entry) {
revoke_entry->trans_id = info->this_trans_id;
return;
}
revoke_entry = jbd_alloc_revoke_entry();
ext4_assert(revoke_entry);
revoke_entry->block = block;
revoke_entry->trans_id = info->this_trans_id;
RB_INSERT(jbd_revoke, &info->revoke_root, revoke_entry);
return;
}
static void jbd_destroy_revoke_tree(struct recover_info *info)
{
while (!RB_EMPTY(&info->revoke_root)) {
struct revoke_entry *revoke_entry =
RB_MIN(jbd_revoke, &info->revoke_root);
ext4_assert(revoke_entry);
RB_REMOVE(jbd_revoke, &info->revoke_root, revoke_entry);
jbd_free_revoke_entry(revoke_entry);
}
}
#define ACTION_SCAN 0
#define ACTION_REVOKE 1
#define ACTION_RECOVER 2
/**@brief Add entries in a revoke block to revoke tree.
* @param jbd_fs jbd filesystem
* @param header revoke block header
* @param recover_info journal replay info*/
static void jbd_build_revoke_tree(struct jbd_fs *jbd_fs,
struct jbd_bhdr *header,
struct recover_info *info)
{
char *blocks_entry;
struct jbd_revoke_header *revoke_hdr =
(struct jbd_revoke_header *)header;
uint32_t i, nr_entries, record_len = 4;
/* If we are working on a 64bit jbd filesystem, */
if (JBD_HAS_INCOMPAT_FEATURE(&jbd_fs->sb,
JBD_FEATURE_INCOMPAT_64BIT))
record_len = 8;
nr_entries = (jbd_get32(revoke_hdr, count) -
sizeof(struct jbd_revoke_header)) /
record_len;
blocks_entry = (char *)(revoke_hdr + 1);
for (i = 0;i < nr_entries;i++) {
if (record_len == 8) {
uint64_t *blocks =
(uint64_t *)blocks_entry;
jbd_add_revoke_block_tags(info, to_be64(*blocks));
} else {
uint32_t *blocks =
(uint32_t *)blocks_entry;
jbd_add_revoke_block_tags(info, to_be32(*blocks));
}
blocks_entry += record_len;
}
}
static void jbd_debug_descriptor_block(struct jbd_fs *jbd_fs,
struct jbd_bhdr *header,
uint32_t *iblock)
{
jbd_iterate_block_table(jbd_fs,
header + 1,
jbd_get32(&jbd_fs->sb, blocksize) -
sizeof(struct jbd_bhdr),
jbd_display_block_tags,
iblock);
}
static void jbd_replay_descriptor_block(struct jbd_fs *jbd_fs,
struct jbd_bhdr *header,
struct replay_arg *arg)
{
jbd_iterate_block_table(jbd_fs,
header + 1,
jbd_get32(&jbd_fs->sb, blocksize) -
sizeof(struct jbd_bhdr),
jbd_replay_block_tags,
arg);
}
/**@brief The core routine of journal replay.
* @param jbd_fs jbd filesystem
* @param recover_info journal replay info
* @param action action needed to be taken
* @return standard error code*/
static int jbd_iterate_log(struct jbd_fs *jbd_fs,
struct recover_info *info,
int action)
{
int r = EOK;
bool log_end = false;
struct jbd_sb *sb = &jbd_fs->sb;
uint32_t start_trans_id, this_trans_id;
uint32_t start_block, this_block;
/* We start iterating valid blocks in the whole journal.*/
start_trans_id = this_trans_id = jbd_get32(sb, sequence);
start_block = this_block = jbd_get32(sb, start);
if (action == ACTION_SCAN)
info->trans_cnt = 0;
else if (!info->trans_cnt)
log_end = true;
ext4_dbg(DEBUG_JBD, "Start of journal at trans id: %" PRIu32 "\n",
start_trans_id);
while (!log_end) {
struct ext4_block block;
struct jbd_bhdr *header;
/* If we are not scanning for the last
* valid transaction in the journal,
* we will stop when we reach the end of
* the journal.*/
if (action != ACTION_SCAN)
if (trans_id_diff(this_trans_id, info->last_trans_id) > 0) {
log_end = true;
continue;
}
r = jbd_block_get(jbd_fs, &block, this_block);
if (r != EOK)
break;
header = (struct jbd_bhdr *)block.data;
/* This block does not have a valid magic number,
* so we have reached the end of the journal.*/
if (jbd_get32(header, magic) != JBD_MAGIC_NUMBER) {
jbd_block_set(jbd_fs, &block);
log_end = true;
continue;
}
/* If the transaction id we found is not expected,
* we may have reached the end of the journal.
*
* If we are not scanning the journal, something
* bad might have taken place. :-( */
if (jbd_get32(header, sequence) != this_trans_id) {
if (action != ACTION_SCAN)
r = EIO;
jbd_block_set(jbd_fs, &block);
log_end = true;
continue;
}
switch (jbd_get32(header, blocktype)) {
case JBD_DESCRIPTOR_BLOCK:
if (!jbd_verify_meta_csum(jbd_fs, header)) {
ext4_dbg(DEBUG_JBD,
DBG_WARN "Descriptor block checksum failed."
"Journal block: %" PRIu32"\n",
this_block);
log_end = true;
break;
}
ext4_dbg(DEBUG_JBD, "Descriptor block: %" PRIu32", "
"trans_id: %" PRIu32"\n",
this_block, this_trans_id);
if (action == ACTION_RECOVER) {
struct replay_arg replay_arg;
replay_arg.info = info;
replay_arg.this_block = &this_block;
replay_arg.this_trans_id = this_trans_id;
jbd_replay_descriptor_block(jbd_fs,
header, &replay_arg);
} else
jbd_debug_descriptor_block(jbd_fs,
header, &this_block);
break;
case JBD_COMMIT_BLOCK:
if (!jbd_verify_commit_csum(jbd_fs,
(struct jbd_commit_header *)header)) {
ext4_dbg(DEBUG_JBD,
DBG_WARN "Commit block checksum failed."
"Journal block: %" PRIu32"\n",
this_block);
log_end = true;
break;
}
ext4_dbg(DEBUG_JBD, "Commit block: %" PRIu32", "
"trans_id: %" PRIu32"\n",
this_block, this_trans_id);
/*
* This is the end of a transaction,
* we may now proceed to the next transaction.
*/
this_trans_id++;
if (action == ACTION_SCAN)
info->trans_cnt++;
break;
case JBD_REVOKE_BLOCK:
if (!jbd_verify_meta_csum(jbd_fs, header)) {
ext4_dbg(DEBUG_JBD,
DBG_WARN "Revoke block checksum failed."
"Journal block: %" PRIu32"\n",
this_block);
log_end = true;
break;
}
ext4_dbg(DEBUG_JBD, "Revoke block: %" PRIu32", "
"trans_id: %" PRIu32"\n",
this_block, this_trans_id);
if (action == ACTION_REVOKE) {
info->this_trans_id = this_trans_id;
jbd_build_revoke_tree(jbd_fs,
header, info);
}
break;
default:
log_end = true;
break;
}
jbd_block_set(jbd_fs, &block);
this_block++;
wrap(sb, this_block);
if (this_block == start_block)
log_end = true;
}
ext4_dbg(DEBUG_JBD, "End of journal.\n");
if (r == EOK && action == ACTION_SCAN) {
/* We have finished scanning the journal. */
info->start_trans_id = start_trans_id;
if (trans_id_diff(this_trans_id, start_trans_id) > 0)
info->last_trans_id = this_trans_id - 1;
else
info->last_trans_id = this_trans_id;
}
return r;
}
/**@brief Replay journal.
* @param jbd_fs jbd filesystem
* @return standard error code*/
int jbd_recover(struct jbd_fs *jbd_fs)
{
int r;
struct recover_info info;
struct jbd_sb *sb = &jbd_fs->sb;
if (!sb->start)
return EOK;
RB_INIT(&info.revoke_root);
r = jbd_iterate_log(jbd_fs, &info, ACTION_SCAN);
if (r != EOK)
return r;
r = jbd_iterate_log(jbd_fs, &info, ACTION_REVOKE);
if (r != EOK)
return r;
r = jbd_iterate_log(jbd_fs, &info, ACTION_RECOVER);
if (r == EOK) {
/* If we successfully replay the journal,
* clear EXT4_FINCOM_RECOVER flag on the
* ext4 superblock, and set the start of
* journal to 0.*/
uint32_t features_incompatible =
ext4_get32(&jbd_fs->inode_ref.fs->sb,
features_incompatible);
jbd_set32(&jbd_fs->sb, start, 0);
jbd_set32(&jbd_fs->sb, sequence, info.last_trans_id);
features_incompatible &= ~EXT4_FINCOM_RECOVER;
ext4_set32(&jbd_fs->inode_ref.fs->sb,
features_incompatible,
features_incompatible);
jbd_fs->dirty = true;
r = ext4_sb_write(jbd_fs->bdev,
&jbd_fs->inode_ref.fs->sb);
}
jbd_destroy_revoke_tree(&info);
return r;
}
static void jbd_journal_write_sb(struct jbd_journal *journal)
{
struct jbd_fs *jbd_fs = journal->jbd_fs;
jbd_set32(&jbd_fs->sb, start, journal->start);
jbd_set32(&jbd_fs->sb, sequence, journal->trans_id);
jbd_fs->dirty = true;
}
/**@brief Start accessing the journal.
* @param jbd_fs jbd filesystem
* @param journal current journal session
* @return standard error code*/
int jbd_journal_start(struct jbd_fs *jbd_fs,
struct jbd_journal *journal)
{
int r;
uint32_t features_incompatible =
ext4_get32(&jbd_fs->inode_ref.fs->sb,
features_incompatible);
features_incompatible |= EXT4_FINCOM_RECOVER;
ext4_set32(&jbd_fs->inode_ref.fs->sb,
features_incompatible,
features_incompatible);
r = ext4_sb_write(jbd_fs->bdev,
&jbd_fs->inode_ref.fs->sb);
if (r != EOK)
return r;
journal->first = jbd_get32(&jbd_fs->sb, first);
journal->start = journal->first;
journal->last = journal->first;
/*
* To invalidate any stale records we need to start from
* the checkpoint transaction ID of the previous journalling session
* plus 1.
*/
journal->trans_id = jbd_get32(&jbd_fs->sb, sequence) + 1;
journal->alloc_trans_id = journal->trans_id;
journal->block_size = jbd_get32(&jbd_fs->sb, blocksize);
TAILQ_INIT(&journal->cp_queue);
RB_INIT(&journal->block_rec_root);
journal->jbd_fs = jbd_fs;
jbd_journal_write_sb(journal);
r = jbd_write_sb(jbd_fs);
if (r != EOK)
return r;
jbd_fs->bdev->journal = journal;
return EOK;
}
static void jbd_trans_end_write(struct ext4_bcache *bc __unused,
struct ext4_buf *buf __unused,
int res,
void *arg);
/*
* This routine is only suitable to committed transactions. */
static void jbd_journal_flush_trans(struct jbd_trans *trans)
{
struct jbd_buf *jbd_buf, *tmp;
struct jbd_journal *journal = trans->journal;
struct ext4_fs *fs = journal->jbd_fs->inode_ref.fs;
void *tmp_data = ext4_malloc(journal->block_size);
ext4_assert(tmp_data);
TAILQ_FOREACH_SAFE(jbd_buf, &trans->buf_queue, buf_node,
tmp) {
struct ext4_buf *buf;
struct ext4_block block;
/* The buffer is not yet flushed. */
buf = ext4_bcache_find_get(fs->bdev->bc, &block,
jbd_buf->block_rec->lba);
if (!(buf && ext4_bcache_test_flag(buf, BC_UPTODATE) &&
jbd_buf->block_rec->trans == trans)) {
int r;
struct ext4_block jbd_block = EXT4_BLOCK_ZERO();
ext4_assert(jbd_block_get(journal->jbd_fs,
&jbd_block,
jbd_buf->jbd_lba) == EOK);
memcpy(tmp_data, jbd_block.data,
journal->block_size);
ext4_block_set(fs->bdev, &jbd_block);
r = ext4_blocks_set_direct(fs->bdev, tmp_data,
jbd_buf->block_rec->lba, 1);
jbd_trans_end_write(fs->bdev->bc, buf, r, jbd_buf);
} else
ext4_block_flush_buf(fs->bdev, buf);
if (buf)
ext4_block_set(fs->bdev, &block);
}
ext4_free(tmp_data);
}
static void
jbd_journal_skip_pure_revoke(struct jbd_journal *journal,
struct jbd_trans *trans)
{
journal->start = trans->start_iblock +
trans->alloc_blocks;
wrap(&journal->jbd_fs->sb, journal->start);
journal->trans_id = trans->trans_id + 1;
jbd_journal_free_trans(journal,
trans, false);
jbd_journal_write_sb(journal);
}
void
jbd_journal_purge_cp_trans(struct jbd_journal *journal,
bool flush,
bool once)
{
struct jbd_trans *trans;
while ((trans = TAILQ_FIRST(&journal->cp_queue))) {
if (!trans->data_cnt) {
TAILQ_REMOVE(&journal->cp_queue,
trans,
trans_node);
jbd_journal_skip_pure_revoke(journal, trans);
} else {
if (trans->data_cnt ==
trans->written_cnt) {
journal->start =
trans->start_iblock +
trans->alloc_blocks;
wrap(&journal->jbd_fs->sb,
journal->start);
journal->trans_id =
trans->trans_id + 1;
TAILQ_REMOVE(&journal->cp_queue,
trans,
trans_node);
jbd_journal_free_trans(journal,
trans,
false);
jbd_journal_write_sb(journal);
} else if (!flush) {
journal->start =
trans->start_iblock;
wrap(&journal->jbd_fs->sb,
journal->start);
journal->trans_id =
trans->trans_id;
jbd_journal_write_sb(journal);
break;
} else
jbd_journal_flush_trans(trans);
}
if (once)
break;
}
}
/**@brief Stop accessing the journal.
* @param journal current journal session
* @return standard error code*/
int jbd_journal_stop(struct jbd_journal *journal)
{
int r;
struct jbd_fs *jbd_fs = journal->jbd_fs;
uint32_t features_incompatible;
/* Make sure that journalled content have reached
* the disk.*/
jbd_journal_purge_cp_trans(journal, true, false);
/* There should be no block record in this journal
* session. */
if (!RB_EMPTY(&journal->block_rec_root))
ext4_dbg(DEBUG_JBD,
DBG_WARN "There are still block records "
"in this journal session!\n");
features_incompatible =
ext4_get32(&jbd_fs->inode_ref.fs->sb,
features_incompatible);
features_incompatible &= ~EXT4_FINCOM_RECOVER;
ext4_set32(&jbd_fs->inode_ref.fs->sb,
features_incompatible,
features_incompatible);
r = ext4_sb_write(jbd_fs->bdev,
&jbd_fs->inode_ref.fs->sb);
if (r != EOK)
return r;
journal->start = 0;
journal->trans_id = 0;
jbd_journal_write_sb(journal);
return jbd_write_sb(journal->jbd_fs);
}
/**@brief Allocate a block in the journal.
* @param journal current journal session
* @param trans transaction
* @return allocated block address*/
static uint32_t jbd_journal_alloc_block(struct jbd_journal *journal,
struct jbd_trans *trans)
{
uint32_t start_block;
start_block = journal->last++;
trans->alloc_blocks++;
wrap(&journal->jbd_fs->sb, journal->last);
/* If there is no space left, flush just one journalled
* transaction.*/
if (journal->last == journal->start) {
jbd_journal_purge_cp_trans(journal, true, true);
ext4_assert(journal->last != journal->start);
}
return start_block;
}
static struct jbd_block_rec *
jbd_trans_block_rec_lookup(struct jbd_journal *journal,
ext4_fsblk_t lba)
{
struct jbd_block_rec tmp = {
.lba = lba
};
return RB_FIND(jbd_block,
&journal->block_rec_root,
&tmp);
}
static void
jbd_trans_change_ownership(struct jbd_block_rec *block_rec,
struct jbd_trans *new_trans)
{
LIST_REMOVE(block_rec, tbrec_node);
if (new_trans) {
/* Now this block record belongs to this transaction. */
LIST_INSERT_HEAD(&new_trans->tbrec_list, block_rec, tbrec_node);
}
block_rec->trans = new_trans;
}
static inline struct jbd_block_rec *
jbd_trans_insert_block_rec(struct jbd_trans *trans,
ext4_fsblk_t lba)
{
struct jbd_block_rec *block_rec;
block_rec = jbd_trans_block_rec_lookup(trans->journal, lba);
if (block_rec) {
jbd_trans_change_ownership(block_rec, trans);
return block_rec;
}
block_rec = ext4_calloc(1, sizeof(struct jbd_block_rec));
if (!block_rec)
return NULL;
block_rec->lba = lba;
block_rec->trans = trans;
TAILQ_INIT(&block_rec->dirty_buf_queue);
LIST_INSERT_HEAD(&trans->tbrec_list, block_rec, tbrec_node);
RB_INSERT(jbd_block, &trans->journal->block_rec_root, block_rec);
return block_rec;
}
/*
* This routine will do the dirty works.
*/
static void
jbd_trans_finish_callback(struct jbd_journal *journal,
const struct jbd_trans *trans,
struct jbd_block_rec *block_rec,
bool abort,
bool revoke)
{
struct ext4_fs *fs = journal->jbd_fs->inode_ref.fs;
if (block_rec->trans != trans)
return;
if (!abort) {
struct jbd_buf *jbd_buf, *tmp;
TAILQ_FOREACH_SAFE(jbd_buf,
&block_rec->dirty_buf_queue,
dirty_buf_node,
tmp) {
jbd_trans_end_write(fs->bdev->bc,
NULL,
EOK,
jbd_buf);
}
} else {
/*
* We have to roll back data if the block is going to be
* aborted.
*/
struct jbd_buf *jbd_buf;
struct ext4_block jbd_block = EXT4_BLOCK_ZERO(),
block = EXT4_BLOCK_ZERO();
jbd_buf = TAILQ_LAST(&block_rec->dirty_buf_queue,
jbd_buf_dirty);
if (jbd_buf) {
if (!revoke) {
ext4_assert(ext4_block_get_noread(fs->bdev,
&block,
block_rec->lba) == EOK);
ext4_assert(jbd_block_get(journal->jbd_fs,
&jbd_block,
jbd_buf->jbd_lba) == EOK);
memcpy(block.data, jbd_block.data,
journal->block_size);
jbd_trans_change_ownership(block_rec,
jbd_buf->trans);
block.buf->end_write = jbd_trans_end_write;
block.buf->end_write_arg = jbd_buf;
ext4_bcache_set_flag(jbd_block.buf, BC_TMP);
ext4_bcache_set_dirty(block.buf);
ext4_block_set(fs->bdev, &jbd_block);
ext4_block_set(fs->bdev, &block);
return;
} else {
/* The revoked buffer is yet written. */
jbd_trans_change_ownership(block_rec,
jbd_buf->trans);
}
}
}
}
static inline void
jbd_trans_remove_block_rec(struct jbd_journal *journal,
struct jbd_block_rec *block_rec,
struct jbd_trans *trans)
{
/* If this block record doesn't belong to this transaction,
* give up.*/
if (block_rec->trans == trans) {
LIST_REMOVE(block_rec, tbrec_node);
RB_REMOVE(jbd_block,
&journal->block_rec_root,
block_rec);
ext4_free(block_rec);
}
}
/**@brief Add block to a transaction and mark it dirty.
* @param trans transaction
* @param block block descriptor
* @return standard error code*/
int jbd_trans_set_block_dirty(struct jbd_trans *trans,
struct ext4_block *block)
{
struct jbd_buf *jbd_buf;
struct jbd_revoke_rec *rec, tmp_rec = {
.lba = block->lb_id
};
struct jbd_block_rec *block_rec;
if (block->buf->end_write == jbd_trans_end_write) {
jbd_buf = block->buf->end_write_arg;
if (jbd_buf && jbd_buf->trans == trans)
return EOK;
}
jbd_buf = ext4_calloc(1, sizeof(struct jbd_buf));
if (!jbd_buf)
return ENOMEM;
if ((block_rec = jbd_trans_insert_block_rec(trans,
block->lb_id)) == NULL) {
ext4_free(jbd_buf);
return ENOMEM;
}
TAILQ_INSERT_TAIL(&block_rec->dirty_buf_queue,
jbd_buf,
dirty_buf_node);
jbd_buf->block_rec = block_rec;
jbd_buf->trans = trans;
jbd_buf->block = *block;
ext4_bcache_inc_ref(block->buf);
/* If the content reach the disk, notify us
* so that we may do a checkpoint. */
block->buf->end_write = jbd_trans_end_write;
block->buf->end_write_arg = jbd_buf;
trans->data_cnt++;
TAILQ_INSERT_HEAD(&trans->buf_queue, jbd_buf, buf_node);
ext4_bcache_set_dirty(block->buf);
rec = RB_FIND(jbd_revoke_tree,
&trans->revoke_root,
&tmp_rec);
if (rec) {
RB_REMOVE(jbd_revoke_tree, &trans->revoke_root,
rec);
ext4_free(rec);
}
return EOK;
}
/**@brief Add block to be revoked to a transaction
* @param trans transaction
* @param lba logical block address
* @return standard error code*/
int jbd_trans_revoke_block(struct jbd_trans *trans,
ext4_fsblk_t lba)
{
struct jbd_revoke_rec tmp_rec = {
.lba = lba
}, *rec;
rec = RB_FIND(jbd_revoke_tree,
&trans->revoke_root,
&tmp_rec);
if (rec)
return EOK;
rec = ext4_calloc(1, sizeof(struct jbd_revoke_rec));
if (!rec)
return ENOMEM;
rec->lba = lba;
RB_INSERT(jbd_revoke_tree, &trans->revoke_root, rec);
return EOK;
}
/**@brief Try to add block to be revoked to a transaction.
* If @lba still remains in an transaction on checkpoint
* queue, add @lba as a revoked block to the transaction.
* @param trans transaction
* @param lba logical block address
* @return standard error code*/
int jbd_trans_try_revoke_block(struct jbd_trans *trans,
ext4_fsblk_t lba)
{
struct jbd_journal *journal = trans->journal;
struct jbd_block_rec *block_rec =
jbd_trans_block_rec_lookup(journal, lba);
if (block_rec) {
if (block_rec->trans == trans) {
struct jbd_buf *jbd_buf =
TAILQ_LAST(&block_rec->dirty_buf_queue,
jbd_buf_dirty);
/* If there are still unwritten buffers. */
if (TAILQ_FIRST(&block_rec->dirty_buf_queue) !=
jbd_buf)
jbd_trans_revoke_block(trans, lba);
} else
jbd_trans_revoke_block(trans, lba);
}
return EOK;
}
/**@brief Free a transaction
* @param journal current journal session
* @param trans transaction
* @param abort discard all the modifications on the block?
* @return standard error code*/
void jbd_journal_free_trans(struct jbd_journal *journal,
struct jbd_trans *trans,
bool abort)
{
struct jbd_buf *jbd_buf, *tmp;
struct jbd_revoke_rec *rec, *tmp2;
struct jbd_block_rec *block_rec, *tmp3;
struct ext4_fs *fs = journal->jbd_fs->inode_ref.fs;
TAILQ_FOREACH_SAFE(jbd_buf, &trans->buf_queue, buf_node,
tmp) {
block_rec = jbd_buf->block_rec;
if (abort) {
jbd_buf->block.buf->end_write = NULL;
jbd_buf->block.buf->end_write_arg = NULL;
ext4_bcache_clear_dirty(jbd_buf->block.buf);
ext4_block_set(fs->bdev, &jbd_buf->block);
}
TAILQ_REMOVE(&jbd_buf->block_rec->dirty_buf_queue,
jbd_buf,
dirty_buf_node);
jbd_trans_finish_callback(journal,
trans,
block_rec,
abort,
false);
TAILQ_REMOVE(&trans->buf_queue, jbd_buf, buf_node);
ext4_free(jbd_buf);
}
RB_FOREACH_SAFE(rec, jbd_revoke_tree, &trans->revoke_root,
tmp2) {
RB_REMOVE(jbd_revoke_tree, &trans->revoke_root, rec);
ext4_free(rec);
}
LIST_FOREACH_SAFE(block_rec, &trans->tbrec_list, tbrec_node,
tmp3) {
jbd_trans_remove_block_rec(journal, block_rec, trans);
}
ext4_free(trans);
}
/**@brief Write commit block for a transaction
* @param trans transaction
* @return standard error code*/
static int jbd_trans_write_commit_block(struct jbd_trans *trans)
{
int rc;
struct ext4_block block;
struct jbd_commit_header *header;
uint32_t commit_iblock;
struct jbd_journal *journal = trans->journal;
commit_iblock = jbd_journal_alloc_block(journal, trans);
rc = jbd_block_get_noread(journal->jbd_fs, &block, commit_iblock);
if (rc != EOK)
return rc;
header = (struct jbd_commit_header *)block.data;
jbd_set32(&header->header, magic, JBD_MAGIC_NUMBER);
jbd_set32(&header->header, blocktype, JBD_COMMIT_BLOCK);
jbd_set32(&header->header, sequence, trans->trans_id);
if (JBD_HAS_INCOMPAT_FEATURE(&journal->jbd_fs->sb,
JBD_FEATURE_COMPAT_CHECKSUM)) {
jbd_set32(header, chksum_type, JBD_CRC32_CHKSUM);
jbd_set32(header, chksum_size, JBD_CRC32_CHKSUM_SIZE);
jbd_set32(header, chksum[0], trans->data_csum);
}
jbd_commit_csum_set(journal->jbd_fs, header);
ext4_bcache_set_dirty(block.buf);
ext4_bcache_set_flag(block.buf, BC_TMP);
rc = jbd_block_set(journal->jbd_fs, &block);
return rc;
}
/**@brief Write descriptor block for a transaction
* @param journal current journal session
* @param trans transaction
* @return standard error code*/
static int jbd_journal_prepare(struct jbd_journal *journal,
struct jbd_trans *trans)
{
int rc = EOK, i = 0;
struct ext4_block desc_block = EXT4_BLOCK_ZERO(),
data_block = EXT4_BLOCK_ZERO();
int32_t tag_tbl_size = 0;
uint32_t desc_iblock = 0;
uint32_t data_iblock = 0;
char *tag_start = NULL, *tag_ptr = NULL;
struct jbd_buf *jbd_buf, *tmp;
struct ext4_fs *fs = journal->jbd_fs->inode_ref.fs;
uint32_t checksum = EXT4_CRC32_INIT;
struct jbd_bhdr *bhdr = NULL;
void *data;
/* Try to remove any non-dirty buffers from the tail of
* buf_queue. */
TAILQ_FOREACH_REVERSE_SAFE(jbd_buf, &trans->buf_queue,
jbd_trans_buf, buf_node, tmp) {
struct jbd_revoke_rec tmp_rec = {
.lba = jbd_buf->block_rec->lba
};
/* We stop the iteration when we find a dirty buffer. */
if (ext4_bcache_test_flag(jbd_buf->block.buf,
BC_DIRTY))
break;
TAILQ_REMOVE(&jbd_buf->block_rec->dirty_buf_queue,
jbd_buf,
dirty_buf_node);
jbd_buf->block.buf->end_write = NULL;
jbd_buf->block.buf->end_write_arg = NULL;
jbd_trans_finish_callback(journal,
trans,
jbd_buf->block_rec,
true,
RB_FIND(jbd_revoke_tree,
&trans->revoke_root,
&tmp_rec));
jbd_trans_remove_block_rec(journal,
jbd_buf->block_rec, trans);
trans->data_cnt--;
ext4_block_set(fs->bdev, &jbd_buf->block);
TAILQ_REMOVE(&trans->buf_queue, jbd_buf, buf_node);
ext4_free(jbd_buf);
}
TAILQ_FOREACH_SAFE(jbd_buf, &trans->buf_queue, buf_node, tmp) {
struct tag_info tag_info;
bool uuid_exist = false;
bool is_escape = false;
struct jbd_revoke_rec tmp_rec = {
.lba = jbd_buf->block_rec->lba
};
if (!ext4_bcache_test_flag(jbd_buf->block.buf,
BC_DIRTY)) {
TAILQ_REMOVE(&jbd_buf->block_rec->dirty_buf_queue,
jbd_buf,
dirty_buf_node);
jbd_buf->block.buf->end_write = NULL;
jbd_buf->block.buf->end_write_arg = NULL;
/* The buffer has not been modified, just release
* that jbd_buf. */
jbd_trans_finish_callback(journal,
trans,
jbd_buf->block_rec,
true,
RB_FIND(jbd_revoke_tree,
&trans->revoke_root,
&tmp_rec));
jbd_trans_remove_block_rec(journal,
jbd_buf->block_rec, trans);
trans->data_cnt--;
ext4_block_set(fs->bdev, &jbd_buf->block);
TAILQ_REMOVE(&trans->buf_queue, jbd_buf, buf_node);
ext4_free(jbd_buf);
continue;
}
checksum = jbd_block_csum(journal->jbd_fs,
jbd_buf->block.data,
checksum,
trans->trans_id);
if (((struct jbd_bhdr *)jbd_buf->block.data)->magic ==
to_be32(JBD_MAGIC_NUMBER))
is_escape = true;
again:
if (!desc_iblock) {
desc_iblock = jbd_journal_alloc_block(journal, trans);
rc = jbd_block_get_noread(journal->jbd_fs, &desc_block, desc_iblock);
if (rc != EOK)
break;
bhdr = (struct jbd_bhdr *)desc_block.data;
jbd_set32(bhdr, magic, JBD_MAGIC_NUMBER);
jbd_set32(bhdr, blocktype, JBD_DESCRIPTOR_BLOCK);
jbd_set32(bhdr, sequence, trans->trans_id);
tag_start = (char *)(bhdr + 1);
tag_ptr = tag_start;
uuid_exist = true;
tag_tbl_size = journal->block_size -
sizeof(struct jbd_bhdr);
if (jbd_has_csum(&journal->jbd_fs->sb))
tag_tbl_size -= sizeof(struct jbd_block_tail);
if (!trans->start_iblock)
trans->start_iblock = desc_iblock;
ext4_bcache_set_dirty(desc_block.buf);
ext4_bcache_set_flag(desc_block.buf, BC_TMP);
}
tag_info.block = jbd_buf->block.lb_id;
tag_info.uuid_exist = uuid_exist;
tag_info.is_escape = is_escape;
if (i == trans->data_cnt - 1)
tag_info.last_tag = true;
else
tag_info.last_tag = false;
tag_info.checksum = checksum;
if (uuid_exist)
memcpy(tag_info.uuid, journal->jbd_fs->sb.uuid,
UUID_SIZE);
rc = jbd_write_block_tag(journal->jbd_fs,
tag_ptr,
tag_tbl_size,
&tag_info);
if (rc != EOK) {
jbd_meta_csum_set(journal->jbd_fs, bhdr);
desc_iblock = 0;
rc = jbd_block_set(journal->jbd_fs, &desc_block);
if (rc != EOK)
break;
goto again;
}
data_iblock = jbd_journal_alloc_block(journal, trans);
rc = jbd_block_get_noread(journal->jbd_fs, &data_block, data_iblock);
if (rc != EOK) {
desc_iblock = 0;
ext4_bcache_clear_dirty(desc_block.buf);
jbd_block_set(journal->jbd_fs, &desc_block);
break;
}
data = data_block.data;
memcpy(data, jbd_buf->block.data,
journal->block_size);
if (is_escape)
((struct jbd_bhdr *)data)->magic = 0;
ext4_bcache_set_dirty(data_block.buf);
ext4_bcache_set_flag(data_block.buf, BC_TMP);
rc = jbd_block_set(journal->jbd_fs, &data_block);
if (rc != EOK) {
desc_iblock = 0;
ext4_bcache_clear_dirty(desc_block.buf);
jbd_block_set(journal->jbd_fs, &desc_block);
break;
}
jbd_buf->jbd_lba = data_iblock;
tag_ptr += tag_info.tag_bytes;
tag_tbl_size -= tag_info.tag_bytes;
i++;
}
if (rc == EOK && desc_iblock) {
jbd_meta_csum_set(journal->jbd_fs,
(struct jbd_bhdr *)bhdr);
trans->data_csum = checksum;
rc = jbd_block_set(journal->jbd_fs, &desc_block);
}
return rc;
}
/**@brief Write revoke block for a transaction
* @param journal current journal session
* @param trans transaction
* @return standard error code*/
static int
jbd_journal_prepare_revoke(struct jbd_journal *journal,
struct jbd_trans *trans)
{
int rc = EOK, i = 0;
struct ext4_block desc_block = EXT4_BLOCK_ZERO();
int32_t tag_tbl_size = 0;
uint32_t desc_iblock = 0;
char *blocks_entry = NULL;
struct jbd_revoke_rec *rec, *tmp;
struct jbd_revoke_header *header = NULL;
int32_t record_len = 4;
struct jbd_bhdr *bhdr = NULL;
if (JBD_HAS_INCOMPAT_FEATURE(&journal->jbd_fs->sb,
JBD_FEATURE_INCOMPAT_64BIT))
record_len = 8;
RB_FOREACH_SAFE(rec, jbd_revoke_tree, &trans->revoke_root,
tmp) {
again:
if (!desc_iblock) {
desc_iblock = jbd_journal_alloc_block(journal, trans);
rc = jbd_block_get_noread(journal->jbd_fs, &desc_block,
desc_iblock);
if (rc != EOK)
break;
bhdr = (struct jbd_bhdr *)desc_block.data;
jbd_set32(bhdr, magic, JBD_MAGIC_NUMBER);
jbd_set32(bhdr, blocktype, JBD_REVOKE_BLOCK);
jbd_set32(bhdr, sequence, trans->trans_id);
header = (struct jbd_revoke_header *)bhdr;
blocks_entry = (char *)(header + 1);
tag_tbl_size = journal->block_size -
sizeof(struct jbd_revoke_header);
if (jbd_has_csum(&journal->jbd_fs->sb))
tag_tbl_size -= sizeof(struct jbd_block_tail);
if (!trans->start_iblock)
trans->start_iblock = desc_iblock;
ext4_bcache_set_dirty(desc_block.buf);
ext4_bcache_set_flag(desc_block.buf, BC_TMP);
}
if (tag_tbl_size < record_len) {
jbd_set32(header, count,
journal->block_size - tag_tbl_size);
jbd_meta_csum_set(journal->jbd_fs, bhdr);
bhdr = NULL;
desc_iblock = 0;
header = NULL;
rc = jbd_block_set(journal->jbd_fs, &desc_block);
if (rc != EOK)
break;
goto again;
}
if (record_len == 8) {
uint64_t *blocks =
(uint64_t *)blocks_entry;
*blocks = to_be64(rec->lba);
} else {
uint32_t *blocks =
(uint32_t *)blocks_entry;
*blocks = to_be32((uint32_t)rec->lba);
}
blocks_entry += record_len;
tag_tbl_size -= record_len;
i++;
}
if (rc == EOK && desc_iblock) {
if (header != NULL)
jbd_set32(header, count,
journal->block_size - tag_tbl_size);
jbd_meta_csum_set(journal->jbd_fs, bhdr);
rc = jbd_block_set(journal->jbd_fs, &desc_block);
}
return rc;
}
/**@brief Put references of block descriptors in a transaction.
* @param journal current journal session
* @param trans transaction*/
void jbd_journal_cp_trans(struct jbd_journal *journal, struct jbd_trans *trans)
{
struct jbd_buf *jbd_buf, *tmp;
struct ext4_fs *fs = journal->jbd_fs->inode_ref.fs;
TAILQ_FOREACH_SAFE(jbd_buf, &trans->buf_queue, buf_node,
tmp) {
struct ext4_block block = jbd_buf->block;
ext4_block_set(fs->bdev, &block);
}
}
/**@brief Update the start block of the journal when
* all the contents in a transaction reach the disk.*/
static void jbd_trans_end_write(struct ext4_bcache *bc __unused,
struct ext4_buf *buf,
int res,
void *arg)
{
struct jbd_buf *jbd_buf = arg;
struct jbd_trans *trans = jbd_buf->trans;
struct jbd_block_rec *block_rec = jbd_buf->block_rec;
struct jbd_journal *journal = trans->journal;
bool first_in_queue =
trans == TAILQ_FIRST(&journal->cp_queue);
if (res != EOK)
trans->error = res;
TAILQ_REMOVE(&trans->buf_queue, jbd_buf, buf_node);
TAILQ_REMOVE(&block_rec->dirty_buf_queue,
jbd_buf,
dirty_buf_node);
jbd_trans_finish_callback(journal,
trans,
jbd_buf->block_rec,
false,
false);
if (block_rec->trans == trans && buf) {
/* Clear the end_write and end_write_arg fields. */
buf->end_write = NULL;
buf->end_write_arg = NULL;
}
ext4_free(jbd_buf);
trans->written_cnt++;
if (trans->written_cnt == trans->data_cnt) {
/* If it is the first transaction on checkpoint queue,
* we will shift the start of the journal to the next
* transaction, and remove subsequent written
* transactions from checkpoint queue until we find
* an unwritten one. */
if (first_in_queue) {
journal->start = trans->start_iblock +
trans->alloc_blocks;
wrap(&journal->jbd_fs->sb, journal->start);
journal->trans_id = trans->trans_id + 1;
TAILQ_REMOVE(&journal->cp_queue, trans, trans_node);
jbd_journal_free_trans(journal, trans, false);
jbd_journal_purge_cp_trans(journal, false, true);
jbd_journal_write_sb(journal);
jbd_write_sb(journal->jbd_fs);
}
}
}
/**@brief Commit a transaction to the journal immediately.
* @param journal current journal session
* @param trans transaction
* @return standard error code*/
static int __jbd_journal_commit_trans(struct jbd_journal *journal,
struct jbd_trans *trans)
{
int rc = EOK;
uint32_t last = journal->last;
struct jbd_revoke_rec *rec, *tmp;
trans->trans_id = journal->alloc_trans_id;
rc = jbd_journal_prepare(journal, trans);
if (rc != EOK)
goto Finish;
rc = jbd_journal_prepare_revoke(journal, trans);
if (rc != EOK)
goto Finish;
if (TAILQ_EMPTY(&trans->buf_queue) &&
RB_EMPTY(&trans->revoke_root)) {
/* Since there are no entries in both buffer list
* and revoke entry list, we do not consider trans as
* complete transaction and just return EOK.*/
jbd_journal_free_trans(journal, trans, false);
goto Finish;
}
rc = jbd_trans_write_commit_block(trans);
if (rc != EOK)
goto Finish;
journal->alloc_trans_id++;
/* Complete the checkpoint of buffers which are revoked. */
RB_FOREACH_SAFE(rec, jbd_revoke_tree, &trans->revoke_root,
tmp) {
struct jbd_block_rec *block_rec =
jbd_trans_block_rec_lookup(journal, rec->lba);
struct jbd_buf *jbd_buf = NULL;
if (block_rec)
jbd_buf = TAILQ_LAST(&block_rec->dirty_buf_queue,
jbd_buf_dirty);
if (jbd_buf) {
struct ext4_buf *buf;
struct ext4_block block = EXT4_BLOCK_ZERO();
/*
* We do this to reset the ext4_buf::end_write and
* ext4_buf::end_write_arg fields so that the checkpoint
* callback won't be triggered again.
*/
buf = ext4_bcache_find_get(journal->jbd_fs->bdev->bc,
&block,
jbd_buf->block_rec->lba);
jbd_trans_end_write(journal->jbd_fs->bdev->bc,
buf,
EOK,
jbd_buf);
if (buf)
ext4_block_set(journal->jbd_fs->bdev, &block);
}
}
if (TAILQ_EMPTY(&journal->cp_queue)) {
/*
* This transaction is going to be the first object in the
* checkpoint queue.
* When the first transaction in checkpoint queue is completely
* written to disk, we shift the tail of the log to right.
*/
if (trans->data_cnt) {
journal->start = trans->start_iblock;
wrap(&journal->jbd_fs->sb, journal->start);
journal->trans_id = trans->trans_id;
jbd_journal_write_sb(journal);
jbd_write_sb(journal->jbd_fs);
TAILQ_INSERT_TAIL(&journal->cp_queue, trans,
trans_node);
jbd_journal_cp_trans(journal, trans);
} else {
journal->start = trans->start_iblock +
trans->alloc_blocks;
wrap(&journal->jbd_fs->sb, journal->start);
journal->trans_id = trans->trans_id + 1;
jbd_journal_write_sb(journal);
jbd_journal_free_trans(journal, trans, false);
}
} else {
/* No need to do anything to the JBD superblock. */
TAILQ_INSERT_TAIL(&journal->cp_queue, trans,
trans_node);
if (trans->data_cnt)
jbd_journal_cp_trans(journal, trans);
}
Finish:
if (rc != EOK && rc != ENOSPC) {
journal->last = last;
jbd_journal_free_trans(journal, trans, true);
}
return rc;
}
/**@brief Allocate a new transaction
* @param journal current journal session
* @return transaction allocated*/
struct jbd_trans *
jbd_journal_new_trans(struct jbd_journal *journal)
{
struct jbd_trans *trans = NULL;
trans = ext4_calloc(1, sizeof(struct jbd_trans));
if (!trans)
return NULL;
/* We will assign a trans_id to this transaction,
* once it has been committed.*/
trans->journal = journal;
trans->data_csum = EXT4_CRC32_INIT;
trans->error = EOK;
TAILQ_INIT(&trans->buf_queue);
return trans;
}
/**@brief Commit a transaction to the journal immediately.
* @param journal current journal session
* @param trans transaction
* @return standard error code*/
int jbd_journal_commit_trans(struct jbd_journal *journal,
struct jbd_trans *trans)
{
int r = EOK;
r = __jbd_journal_commit_trans(journal, trans);
return r;
}
/**
* @}
*/